Total de registros: 541909
Total de columnas: 15
Memoria utilizada: 208.27 MB
| # | Columna | No Nulos | Tipo de Dato |
|---|---|---|---|
| 0 | InvoiceNo | 541909 / 541909 | object |
| 1 | StockCode | 541909 / 541909 | object |
| 2 | Description | 540455 / 541909 | object |
| 3 | Quantity | 541909 / 541909 | int64 |
| 4 | InvoiceDate | 541909 / 541909 | datetime64[ns] |
| 5 | UnitPrice | 541909 / 541909 | float64 |
| 6 | CustomerID | 541909 / 541909 | float64 |
| 7 | Country | 541909 / 541909 | object |
| 8 | TotalVentas | 541909 / 541909 | float64 |
| 9 | Year | 541909 / 541909 | int32 |
| 10 | Month | 541909 / 541909 | int32 |
| 11 | Quarter | 541909 / 541909 | int32 |
| 12 | DayOfWeek | 541909 / 541909 | int32 |
| 13 | Hour | 541909 / 541909 | int32 |
| 14 | YearMonth | 541909 / 541909 | object |
El dataset ha sido dividido en:
Se mantiene el orden temporal para evitar fuga de información.
| Modelo | MAE | RMSE |
|---|---|---|
| ARIMA | 24632.72 | 29086.40 |
| Regresión Lineal | 28453.76 | 32578.18 |
| Random Forest | 34275.41 | 38779.61 |
| Prophet | 77749.40 | 80210.21 |
| LSTM | 30459.05 | 33871.77 |
El modelo más preciso basado en RMSE Test es:
Este modelo ha demostrado ser el más preciso para predecir ventas futuras.
| Quantity | UnitPrice | CustomerID | TotalVentas | |
|---|---|---|---|---|
| count | 541909.000000 | 541909.000000 | 406829.000000 | 541909.000000 |
| mean | 9.552250 | 4.611114 | 15287.690570 | 17.987795 |
| std | 218.081158 | 96.759853 | 1713.600303 | 378.810824 |
| min | -80995.000000 | -11062.060000 | 12346.000000 | -168469.600000 |
| 25% | 1.000000 | 1.250000 | 13953.000000 | 3.400000 |
| 50% | 3.000000 | 2.080000 | 15152.000000 | 9.750000 |
| 75% | 10.000000 | 4.130000 | 16791.000000 | 17.400000 |
| max | 80995.000000 | 38970.000000 | 18287.000000 | 168469.600000 |